The Isolation of an Octulose and an Octitol from Natural Sources: D

May 1, 2002 - ... D-glycero-D-manno-Octulose and D-erythro-D-galacto-Octitol from the Avocado and D-glycero-D-manno-Octulose from Sedum Species1,2...
0 downloads 0 Views 1MB Size
Download PDF
342s

,I.J. CIL~RLSON AND NELSONK. IIICIIT~IYER

Anal. Calcd. for C d L O i : C, 69.94; H , -1.9i. Fouud: C, 69.76; H, 5.11.

Acknowledgments.-1T-e wish to thank Dr. D. L. NacDonald of this Laboratory for performing the

l70L 81

periodate oxidations. Analyses were carried out by the Institutes' Analytical Services Unit under the direction of Dr. W. C. Alford. BETHESDA14, MD.

[COSTRIRL-TIOS F R O M THE SATIOSAL INSTITUTE OF 'ARTHRITIS ASD METABOLIC HEALTH]

DISEASES,NATIOSAL

ISSTITUTES O F

The Isolation of an Octulose and an Octitol from Natural Sources : D-glycero-D-mannoOctulose and D-euyfhro-D-galacto-Octito1 from the Avocado and D-glycero-D-mannoOctulose from Sedum Species'J B\' -1.J. C H X R L S O K 3 AND NELSOXK. RICHTVYER RECEIVED SOVEMBER 30, 1959 From the avocado, a long-known source of D-manno-heptulose and perseitol (D-glycero-D-galacto-heptitol), we have isolated small amounts of D-glycero-D-manno-octulose and D-erythro-o-galacto-octitol,and have obtained strong evidence for the presence a l x of D-talo-heptulose. From Sedum species, a well-known source of sedoheptulose (D-altro-heptulose), we have isolated small amounts of the same o-glycero-o-manno-octulose and also of 8-sedoheptitol (D-g~yCerO-D-glucO-heptitO~~. We have established the structures of the first known naturally occurring octulose and octitol by degradation studies and by t h e cyanohydrin synthesis from D-glycero-D-manno-heptose. CHzOH CH20H Sature has given us two heptitols and two hepI I tuloses that have been known for a long time: HCOH C=O perseitol( D-g/ycPro-D-gdacto-heptitol, I) from the avocado4; voleinitol (D-glycero-D-manno-heptitol, I 111), discovered in the mushroom Lactarius BoleI HOfH Inus Fr.j and found later in the roots of several HOCH I species of Primula,6 in lichen^,^ and in brown and HCOH HCOH red algae8; D-manno-heptulose (11) from the avocadog; and sedoheptulose (D-aho-heptulose, IV), I I discovered in Sedum spectabile Bor.l0 and in other CHLOH CHzOH Sedum species, where it accumulates to the extent Perseitol D-mannoof about 1%, and found later to be widely distrib(I) Heptulose uted in nature (though usually in very small (11) amounts) and to be a fundamental factor in plant and animal metabolism.l' Thus, perseitol (I) CHzOH CHiOH I and D-manno-heptulose (11) occur .together in the HCOH C=O avocado, and, in 1951, Kordal and Oisethl? isolated I both volernitol (111) and sedoheptulose (IV), the HOCH HOCH latter being identified as the crystalline di-0I benzylidene derivative of sedoheptulosan, from the I I dried root stock of Primula elatior (L) Hill. We HCOH HCOH wish to announce that a third naturally occurring I I combination of heptitol and heptulose has now been HCOH HCOH I I established by our discovery that P-sedoheptitol HCOH HCOH (D-glycero-D-gluco-heptitol, v ) accompanies sedo(1) Presented in part before t h e Division of Carbohydrate Chemistry a t t h e Atlantic City Meeting of t h e American Chemical Society, September 16, 1959. (2) A preliminary communication appeared in THISJOURIAL, 81, 1512 (1959). (3) Visiting Scientist of t h e Public Health Service, April, 1958, t o Xovember, 1959. (4) J. B. Avequin, J . chim. mid., pharm. loxicol., [ l ] 7 , 467 (1831); see R. M. Hann and C. S. Hudson, TEIS J O U R N A L , 61, 338 (1939). ( 5 ) E. Bourquelot, Bull. soc. mycoiogique France, 6, 132 (1889). (6) J. Bougault and G. Allard, Compl. vend., 186, 796 (1902). (7) Y . Asahina and hl. Kagitani, Be?., 67, 804 (1934); B. Lindberg, A. Misiorny and C. .4. Wachtmeister. Acta Chem. Scami., 7, 591 (1953), B. Lindberg, i b i d . , 9 , 917 (19%). ( 8 ) B. Lindberg and J. Paju, Acla Chein. Scond., 8, 817 (1954); B. Lindberg, ibid., 9, 1097 (1955). F. B. LaForge, J . B i d . Chent., 26, 511 (1917). (10) F. B. LaForge and C. S. Hudson, ibid., 8 0 , 61 (1917). (11) For recent reviews see, for example, A. S o r d a l , Tidsskr. Kje?ni, Bevguesen M e t . , 6, 77 (1956); A. Bonsignore, M. Orunesu, S . Pontremoli a n d P. Vegetti, Giorn. biochim., 6, 203 (1956) (12) A. Sordal and D. oiseth, Acto Chrnl. Scuwd,, 6, 1289 (1961).

I

I

CH2OH

I

I

I

HOCH

C=O

I

I HCOJI

HCOH

HCOH 1

I I CH2OH CHzOH SedoVolemitol heptulose (111) CHzOH

I

I

HCOH

CHO

HCOH

I

HOCH

(TTII)

(IX)

HOCH HCOH

(XI

I I

HCOH

HCOH

HCOH

HCOH

HCOH

HCOH

HcOH

HCOH

H h H

+OH I

I

HOCH

CHO

HCOH

I

HCOH

COOH r-C=O

HCOH

I

1

HCOH

I CH~OH CH~OH D-erythro- r-glycero- D-threo-LD-inannogalactoD-taloOctulose Octow Octitol

LH~OH

HYOH

I

HOCH

I

HCOH

I

I CH20H p-Sedoheptitol

CH3OH

I HCOH I

HCOH

HCOH I

(VI

HOAH I

I

HCOH

HCOH

(1x7)

C=O

HOCH

I

I

HCOH

I

I

HOCH

CHzOH

I

HCOH

I

CH~OH CH~OH D-eiythro- D-glyceroD-gdacto- n-mannoOctitol Octulose (171 1 (S'II) COOH

CHcOH

I

I

I

I I HCOH I I

CHzOH

CHzOH

CHzOH

CH20H

CHJOH

XI

XI1

SI11

XlV

XV

July 5, 1960

OCTULOSE AND OCTITOL FROM AVOCADO AND SEDUX

3429

showed [a]2 0 ~ 4-2 O in chloroform. Inasmuch as the reduction of D-manno-heptulose (11) furThe occurrence of more than one heptulose in nishes the two alditols perseitol (I) and volernitol the same plant material has been suggested by (111),both of which are found in nature, and sedoNordal and Oiseth,12 by Williams and Bevenue,13 heptulose (IV) furnishes volemitol (111) and @and by Williams, Potter and Bevenue.14 Their sedoheptitol (V), we reasoned that the avocado evidence was based only on the number of spots octitol should be related to the avocado octulose on paper chromatograms that were revealed with (VII) and therefore should be either D-erythro-D(VIII). orcinol-acid sprays, which we now know to be galacto-octitol (VI) or D-erythro-D-talo-octitol capable of detecting the non-reducing anhydro- A test oxidation with -4cetobacter suboxyhns heptuloses as well. However, we have isolated showed that the avocado octitol was oxidized to from the avocado, through extensive fractionation, an octulose that migrated on a paper chromatogram a small amount of a heptulose that is indistinguish- more slowly than the avocado octulose (VII); able from D-talo-heptulose on paper chromatograms from this we concluded that the avocado octitol and that yielded, on degradation with oxygen in could not be D-erythro-D-talo-octitol (VIII) for cold alkaline solution, l5 a lactone that was indis- VI11 could have been oxidized only to the avocado tinguishable from D-talonic lactone on paper octulose (VII). On the other hand, rearrangement chromatograms. While we cannot consider the of the known D-threo-L-galacto-octose (X) in proof to be conclusive until we obtain more ma- boiling pyridinelYafforded a solution that contained terial and isolate the heptulose in crystalline form, an octulose that migrated on paper chromatowe believe that we have a t least strong evidence grams, run in four different solvent systems, a t the same rate as the octulose produced by the for the occurrence of D-talo-heptulosein the avocado. Much more interesting is our discovery of an action of Acetobacter suboxydans on the avocado octulose and a closely related octitol in the avocado. octitol. Since the ketose expected from the pyriThe octulose was detected in the aqueous alcoholic dine rearrangement of the aldose X would be extract of the Californian avocado (Calavo, D-glycero-L-manno-octulose, and the ketose exFuerte variety) when its paper chromatogram, pected from the biochemical oxidation of D-erythrosprayed with orcinol-hydrochloric acid and heated D-galacto-octitol (VI) a t C7 would be L-glycero-Din an oven a t l l O o , revealed a crimson-colored manno-octulose (IX), these two octuloses would spot that faded rapidly to gray; a similar spot be enantiomorphs, and that conclusion was supwas obtained with a synthetic octulose. Heptu- ported by their identical Rf-values. Accordingly, loses, by contrast, give blue or bluish-green spots the mixture of components from the pyridine under these conditions. The avocado extract rearrangement was reduced with sodium borohywas freed from gums by precipitation with meth- dride and the alditols were separated readily into anol, deionized, fermented, oxidized with hypo- two fractions, the larger one being the known bromite to remove aldoses, and the mixture of D-threo-L-galacto-octitol's and the smaller one being polyhydric alcohols and ketoses separated and the new D-threo-L-talo-octitol ( = L-erythro-L-galactopurified by a combination of crystallization and octitol). The latter was proved by its physical cellulose column chromatography. The alditol properties and those of its octaacetate to be the fractions yielded myo-inositol and perseitol in enantiomorph of the avocado octitol, which must addition to the octitol to be described later in this therefore be D erythro-D-galacto-octitol (VI). paper. Finally, the synthesis of the avocado octitol The octulose (1 g. from 27 kg. of avocado pulp) was undertaken, starting with some D-glycerowas a sirup, with [ a I z 0+~ Z o o in methanol. As D-manno-heptose (XI) that was available from stated in our preliminary communication, it earlier studies in this Laboratory.20 The addition was characterized by a crystalline 2,5-dichloro- of hydrogen cyanide to this heptose and subsequent phenylhydrazone; on oxidation with two molar hydrolysis of the nitrile mixture yielded a crystalequivalents of lead tetraacetate according to the line octonic acid; this was converted to a crystalmethod of Perlin and Brice,16 followed by acid line lactone, and this in turn to a crystalline phenylhydrolysis, i t yielded D-ribose; and on degradation hydrazide whose dextrorotation indicated that the with oxygen in cold alkaline s o l ~ t i o n 'i~t yielded hydroxyl group a t C2 was on the right and that D-glycero-D-manno-heptonic lactone. On the basis the principal product in this synthesis was the of these degradations we concluded that the octu- D-erythro-D-galacto-octonicacid (XII) rather than lose must be D-glycero-D-manno-octulose (VII). the epimeric acid XIII. Reduction of a portion The octitol from the avocado was obtained as a of the D-erythro-D-galacto-octonicy-lactone (XIV) monohydrate that melted a t 169-170 O and showed with excess sodium borohydride yielded D-erythro[a]2 0 ~- 11O in 5% aqueous ammonium molyb- D-galacto-octitol (VI) identical with the octitol date and +66" in the acidified molybdate solu- isolated from the avocado. tion." Its octaacetate melted at 99-100' and Reduction of the remainder of the lactone with sodium amalgam afforded D-erythro-D-galacto-oc(13) K. T. Williams and A. Bevenue, J . Assoc. O$c. Agr. Chemists, 34, 817 (1951). tose ( x v ) ; its mutarotation upward to [CY]"".

heptulose (IV) in the leaves and stems of Sedum

spectabile.

(14) K. T. Williams, E. F. Potter and A. Bevenue, ibid., 31, 483 (1952). (15) T h e method of 0. Spengler and A. Pfannenstiel, Z.Wirfschafisgru$pe Zuckerind., 81, Tech. TI. 547 (1935). (16) A. S.Perlin and C . Brice, Can. J . Chem., 34, 541 (1986). (17) PI' K. Richtmyer and C. S. Hudson, TRISJOURNAL, 73, 2249

(1951).

(18) W D. Maclay, R M Hann and C. S Hudson, ibtd., 60, 1035

(1938).

(19) T h e method of S. Danilow, E. Venus-Danilowa and P. Schantarovitsch, Bcr., 6.3, 2269 (1930). (20) D. A. Rosenfeld, N. K Richtmyer and C. S. Hudson, THIS JOURAAL, 73, 4907 (1951).

3430

A. J. CIIARLSON AND EELSON K. RICHTMYER

4-64' in water indicated the crystalline material to be the B-modification of the sugar. Rearrangement of this octose in pyridine resulted in the formation of an octulose that migrated on paper chromatograms in four solvent systems a t the same rate as the avocado octulose (VII). Unfortunately there was not enough of the octose to rearrange and then attempt to isolate and identify the octulose with that from the avocado. Instead, the synthetic octose was converted to its phenylosazone and the latter proved to be identical with that derived from the avocado octulose through melting points, mixed melting points, infrared spectra and X-ray powder diffraction patterns. Having discovered an octulose in the avocado, which is a good source of D-manno-heptulose, we reinvestigated the principal source of the other known naturally occurring heptulose and detected an octulose also in extracts of two Sedum species. We, therefore, attempted to isolate the Sedum octulose and, after extensive treatment and fractionation of the complex mixture of substances in the aqueous extracts of Sedum, succeeded in obtaining 0.16 g. of essentially pure octulose from 130 kg. of Sedum. The rotation of the sirupy octulose, [ a ]z o +20° ~ in methanol, was the same as that of the avocado cctulose. The two octuloses ran a t the same rate on paper chromatograms in four different solvent systems, and their infrared spectra were identical. Degradation of the Sedum octulose with two molar equivalents of lead tetraacetate and with oxygen in cold alkaline solution yielded the same substances, as determined by paper chromatography, that we had isolated and characterized from the larger-scale degradations of the avocado octulose. Finally, the phenylosazone of the Sedum octulose was shown to be identical with those of the avocado octulose and the synthetic D-erythro-D-gakcto-octose. Thus we have shown that both the avocadoz1 and Sedum contain D-glycero-D-inanno-octulose (VII) and the avocado contains the closely related D-erythro-D-galacto-octitol(VI). We may predict that the same octitol will be found, though in very small amounts, also in Sedum. During the course of fractionation of our Sedum extracts we have isolated D-mannitol and myo-inositol, which have frequently been found in plant extracts. We have isolated, for the first time from a plant source, also P-sedoheptitol (V), as mentioned earlier. This substance is related to sedoheptulose (IV) in the same way that perseitol (I) is related to D-manno-heptulose (11) and the avocado octitol (VI) to the octulose (VII), that is, by the addition of hydrogen to the carbonyl group in such a way that the hydroxyl group introduced a t C2 is trans to the one already present a t C3. In regard to the biosynthesis of these 7 - and 8-carbon ketoses and alditols22we might note that in addition to the earlier-known heptulose syntheses (21) Paper chromatograms of old mother liquors from the unfermented extracts of Florida avocados also showed t h e octulose spot, and so t h e octulose does not appear t o be a constituent t h a t is characteristic only of Calavo 1-arieties. However, t h e relative amounts of octulose may change with the variety, a s we know t o be true of t h e Dnianno-heptulose content; see reference 9; E. 11. Alontgomery and C. S. Hudson, THISJOURNAL, 61,1654 (1939); and A. Nordal and .4. A. Benson, ibid., 76, 5054 (1954).

VOl. 82

through the action of aldolases, Racker and Schroederz3 have shown recently that an octulose 8phosphate is formed when D-ribose 5-phosphate and D-fructose 6-phosphate are incubated in the presence of a transaldolase, and that Sephton and Jonesz4 have incubated D-fructose 1,6-diphosphate with each of the four pentoses D-ribose, D-lyxose, D-xylose and L-arabinose in the presence of purified rabbit muscle aldolase and isolated four octuloses, each with a D-threo conl'lguration a t C3 and C4 of the ketose formed. We are continuing our investigations on the carbohydrate constituents of the avocado and Sedum, particularly in an attempt to isolate additional higher-carbon sugars and alditols.

Experimental Paper chromatograms were run on Whatman KO. 1 filter paper by the descending method using solvent A , l-butanolethanol-water (40: 11 : 19); solvent B, ethyl acetate-acetic acid-formic acid-water (18:3 : 1:4); solvent C, l-butanolpyridine-water (6: 4:3), or solvent D, 2-butanone-acetic acid-water saturated with boric acid (9: 1 : 1). Spray reagents used were aniline hydrogen phthalate for aldoses, orcinol-hydrochloric acid or orcinol-trichloroacetic acid for ketoses, p-anisidine hydrochloride for aldoses and ketoses, alkaline hydroxylamine-ferric chloride for lactones, and ammoniacal silver nitrate or sodium metaperiodate-potassium permanganate for sugars and alditols. The R,,,,, and Rleotone refer t o the rate of movevalues for R,,,,, ment of the compounds on paper chromatograms relative to that of sucrose, perseitol and D-glyCevo-D-nznnno-h~~t[~tliC lactone, respectively. .ill concentrations were carried o u t i n naczio a t temperatures not over 50". Products from the Avocado. Isolation.-The crushed pulp from 94 ripe avocados (Calavo, Fuerte variety), weighing 27 kg., was extracted with 20y0 ethanol according to the general procedure of Montgomery and Hudson.2s T h e extract was concentrated to 2 1. and the gums were precipitated by pouring the solutiotl into 16 1. of methanol. The granular product was filtered, dissolved in 2 1. of water, and reprecipitated in the same manner. The combined methanol filtrates mere concentrated and an aqueous solution of the residue was deionized by passage through columns of A4mberlite IR-120 and Duolite A-4 ion-exchange resins. The solution was concentrated to 400 ml., 20 g. of D-glucose and 3 cakes of bakers' yeast were added, and fermentation was allowed to proceed for 45 hours a t 37". The solution was then filtered, deionized, and Concentrated to a sirup that yielded readily, when dissolved in methanol, 1.1 g. of crystalline materid. X paper chromatogram, run for 72 hours in solvent A nn(1 sprayed with ammoniacal silver nitrate, indicated that this material was a mixture of perseitol, the octitol (1'1, R,,,. 0.83),and myo-inositol (Ruere0.45), whose separation and identification will be described below. The filtrate from the 1.1g. of polyols was concentrated to a sirup. Paper chromatograms, run in solvents A , B and C, and sprayed with aniline hydrogen phthalate, indicated the presence of arabinose and possibly xylose. Accordingly, all aldoses were removed by dissolving the sirup in 350 ml. of water and oxidizing with 4 ml. of bromine in the presence of 3 g. of barium benzoate for 20 hours in the dark. Excess bromine was removed by aeration, precipitated benzoic acid by filtration, and the solution deionized with Amberlite III120 and Duolite -4-4 resins. Concentration of the solution left 16 g. of sirup that appeared to consist entirely of ketoses and alditols. Fractionation of this sirup was carried out on a cellulose column 325 mm. high and 50 mm. in diameter. Thus were obtained sirup X (6.4 g.) by elution with I-butanol (22) Cf.L. Hough and J. K. N. Jones, Admnces i n Carbohydrale Chcm., 11, 185 (1956). (23) E. Kacker a n d E. Schroeder. AYch. Biochern. Biophys., 66, 241 (1957).

(24) H . H . Sephton and J. K. N. Jones, Abstracts of Papers, Atlantic City Meeting of t h e American Chemical Society, Sept. 13-18, 1969, p. 14E,and personal communications from those authors. (25) E. AI. Montgomery and C. S. Hudson, THISJ O U R N A L , 61, lt734 (1939).

July 5, 1960

OCTULOSEAND OCTITOL FROM AVOCADO AND SEDUM

half-saturated with water; sirup B (3.9 9.) by elution with 1-butanol three-quarters saturated with water; and finally sirup C (3.0 9.) by elution with 75% aqueous methanol. Sirup A. Isolation of Dtalo-Heptulose.-Examination of this sirup by paper chromatography indicated the presence of -fructose, D-manno-heptulose and a second heptulose that ran somewhat faster than D-manno-heptulose. In order to purify the second heptulose, sirup A was fractionated on the cellulose column using 1-butanol-water (14: 1) and 4.1 g. of sirup was removed; this sirup was refractionated using 1-butanol-water (28: 1) and 2.2 g. of sirup obtained that still contained small amounts of &fructose and D-manno-heptulose. Of this, 1.2 g. was fractionated on a cellulose column using benzene-methanol-water (300: 100:4.) The 0.6 g. of sirupy heptulose thus obtained appeared to be free from Dmanno-heptulose but still contained some D-fructose. Most of the latter was then removed by fermentation. Because the final sirup (0.48 g.) was believed to be principally sedoheptulose, it was dissolved in 25 ml. of 0.5 N sulfuric acid; ~ on however, its rotation did not change from [ a I z 0+2' standing 48 hours a t 20'. The solution was heated 2 hours on the steam-bath, deacidified with Duolite A-4, and concentrated to a sirup (0.44 g.). Paper chromatographic examination in all four solvent systems indicated that the major component was not sedoheptulosan but that it might be Dtalo-heptulose (pure: [aImD f12.9026)containing small amounts of D-fructose and probably non-reducing anhydroheptuloses. Because the sirup could not be induced to crystallize when inoculated with authentic D-tdo-heptulose, a 34-mg. sample of it was degraded by bubbling oxygen through its ice-cold solution in N potassium hydroxide for 20 hours. After removal of cations with Amberlite IR-120 and concentration, the residue was examined by paper chromatography in all four solvent systems; the major component was a lactone that ran a t the same rate as D-talonic lactone. Sirup B. Isolation of the Octitol VI and Characterization as Derytkro-Dgalacto-Octitol 0ctaacetate.-A solution of sirup B in ethanol containing a small amount of water deposited 0.95 g. of needles melting a t 166-168'. Paper chromatography in solvent A identified the material as the octitol (R,,,. 0.83) contaminated with a very small amount of perseitol. The acetylation of 0.3 g. of this material by two successive treatments with acetic anhydride and pyridine yielded a sirup that crystallized. Recrystallization from ethanol afforded 0.35 g. of the octitol octaacetate as chunky prisms melting a t 99-100' and showing [a]% +a0 in chloroform (c 3.4). Anal. Calcd. for C2,H81O16: C, 49.83; H , 5.92; CHr CO, 59.5; mol. wt., 578. Found: C, 49.74; H, 5.88; CHsCO, 59.4; mol. wt., 570 (Signer method: isothermal distillation in acetone). Sirup C. Isolation of myo-Inositol.-A solution of sirup C in ethanol deposited 0.4 g. of needles with Rperm0.45 in solvent A (chromatogram sprayed with ammoniacal silver nitrate). The product was identified as myo-inositol through direct comparison on paper chromatograms, identical infrared spectra (Nujol mull), and melting point and mixed melting point of 228-229'. Confimatory proof was obtained by acetylation to myo-inositol hexaacetate, whose melting point of 219-220' was not depressed when the substance was mixed with an authentic sample. Purification of o-erythro-D-galacto-Octitol (VI).-The 1.1 g. of crystalline, mixed alditols that had separated before the bromine oxidation and subsequent fractionation on a cellulose column was combined with the mother liquor remaining after crystallization of the myo-inositol from sirup C, and the concentrated product was fractionated on the cellulose column. Elution with 1-butanol half-saturated with water followed by 1-butanol three-quarters saturated with water yielded, from the latter fractions, 1.2 g. of chromatographically pure octitol. The Derythro-D-galacto-octitol monohydrate was recrystallized from aqueous ethanol, separating as elongated prisms melting at 169-170'. It showed [a] m~ -11' in 5y0 aqueous ammonium molybdate (c 0.40) and t-66' in the acidified molybdate solution (c 0.32).'7 Recrystallization from methanol containing only a small amount of water furnished the anhydrous octitol with the same m.p. of 169-170'. (26) J. W. Pratt and N. K. Richtmyer, TmS JOURNAL, 17, C326 (1955).

3431

A d . Calcd. for C&IlsO~H~O: C, 36.92; H, 7.75. Found (hydrate): C, 37.01; H , 7.55. Calcd. for CsHuOs: C, 39.67; H, 7.49. Found (anhydrous): C, 39.51; H, 7.80. Isolation of D-glycero-D-manno-Octose (VII) from Sirup B.-After removal of octitol from the ethanol solution of sirup B as described above, the mother liquor was concentrated and examined by paper chromatography. When the chromatogram was sprayed with orcinol-hydrochloric acid and heated at l l O o , a crimson-colored spot developed (R.,,, 1.2 in solvent B and 2.6 in solvent D) and then faded to gray. This behavior is believed to be characteristic of an octulose, and the same color sequence was observed when a synthetic octulose was treated similarly. Fractionation of the remainder of sirup B was carried out on the cellulose column. Elution with benzene-methanol-water (200: 100:6) yielded a fraction weighing 1.2 g., with [ C Y ] % 4-20" in methanol ( c 1.1). This sirup appeared t o be essentially homogeneous and free from alditols when examined in solvent D, which is capable of separating the octulose from the octitol. For characterization a small portion of it was converted, by the procedure described by Mandl and Neuberg,n to Dglycero-Dmanno-octulose 2,5-dichlorophenylhydrazone. The product was recrystallized from 40% aqueous ethanol as very pale yellow needles melting a t 169-170". Anal. Calcd. for CllH&12N20,: C1, 17.76. Found: C1, 17.59. Degradation of Dglycero-Dmanno-Octose (VII) to D Ribose with Lead Tetraacetate.-To a solution containing 147 mg. of the sirupy octulose in 80 ml. of glacial acetic acid and 1.6 ml. of water was added 550 mg. (2.0 molar equivalents) of lead tetraacetate. After 15 minutes a 10% solution of oxalic acid in glacial acetic acid was added dropwise until no more lead oxalate precipitated. The solution was a t e r e d and then concentrated to remove most of the acetic acid. The residue was dissolved in 30 ml. of water and the solution heated on the steam-bath for 9 hours to effect complete hydrolysis of formyl and glycolyl groups. The acids were removed by passage through a column of Duolite A-4 ion-exchange resin and the solution was evaporated to yield 72 mg. of sirup with [ C Y ] ~ ~-1 D0' in water (Dribose shows Paper chromatograms in all 4 solvent sys[~]*OD. -20'). tems mdicated that ribose was the major component of the smp. It was identified as Dribose by conversion t o its crystalline fi-tolylsulfonylhydrazone (30 mg.) with melting point 158-159' decomp. and [ U ] ~+17' D in dry pyridine (c 0.3); the reported values" are 164' and +14", respectively. A mixture with authentic material melted also at 158-159'' dec Degradation of Dglycero-wnnnno-Octose (VII) to D glycero-D-manno-Heptonic Lactone with Oxygen in Alkaline Solution.-A 210-mg. portion of the octulose sirup was dissolved in 15 ml. of N potassium hydroxide that had previously been cooled to 0' and saturated with oxygen. Oxygen was bubbled through the solution for 16 hours a t 0' and then for 8 hours a t 20-25'. The solution was freed of cations with Amberlite IR-120 and concentrated to a sirup. Paper chromatograms run in solvents A, B and C indicated that a lactone with Rtmtcns = 1.0 was the major component. The material was chromatographed on a cellulose column and the lactone eluted with 1-butanol half-saturated with water. The appropriate fractions were combined and concentrated to a sirup that yielded, with 95% ethanol, fine needles of D glycero-D-manno-heptonic lactone. Upon recrystallization from 95% ethanol the 15 mg. of product melted at 164-165' to a thick sirup; this value was undepressed when the substance was mixed with freshly recrystallized authentic lactone.2o The rotation, [ C Y ] ~+48' D in water (c 0.2), was similar to the rotation +40.4' reported earlier. Infrared spectra of the two lactones in Nujol mull were identical. Oxidation of oeryt~ro-Dgalacto-Octitol(VI) by Acetobacter suboxyduns.-A 100-mg. portion of the octitol from the avocado was dissolved in 10 ml. of distilled water containing 50 mg. of Difco yeast extract, 30 mg. of potassium dihydrogen phosphate and 5 mg. of Dglucose. The solution was sterilized in the autoclave and inoculated with 0.4 ml. of a culture of A . suboxydunszg grown on a yeast extract and D-

.

(27) I. Mandl and C. Neuberg, Arch. Biochem. Biophys., 86, 326

(1952). (28) D. 0. EPrtuby, L. Hough and I. K. N. Jones, J . Chem. Soc., 3416 (1951). (29) American Type Culture Collection No. 621.

3432

A.J. CHARLSW AXD NELSOXK. RICIITMYER

I'cll, s2

glucose medium. ;ifter 6 days at 30' the solution was consist of the lactone free from the acid as determined by diluted with 35 nil. of methanol, filtered, deionized with paper chromatography. The lactone had Rlaotone0.83 in .'imberlite I R-120 and Duolite A-4 ion-exchange resins, and solvent B and 1.1 in solvent D. Recrystallization from concentrated. Paper chromatograms sprayed with the ethanol containing a small amount of water afforded 0.47 g. orcinol-hydrochloric acid reagent indicated the presence of of n-erythro-~-galacto-octonic?-lactone. The product apan octulose through a typical crimson spot that faded peared to contain some solvent of crystallization and so i t was rapidly to gray. I n solvents B and D the octulose moved dried for 8 hours in vacuo at 100". It then melted a t 152more slowly than the naturally occurring D-glycero-D-man153' and showed [al2'3~ -66" in water ( c 1). Because the no-octulose (VII). However, in solvents A, B, C and D the lactone was obtained only at a relatively high temperature i t octulose produced by A . suboxydansmoved a t the same rate must be considered a ?-lactone. as the D-glycevo-L-manno-octulose produced in the pyridine .1naZ. Calcd. for C ~ H I ~ O C,~40.34; : H, 5.92. Found: rearrangement of D-thveo-L-galacto-octose (X), as described C, 40.70; H, 5.66. in the following paragraph. The octulose produced by bioThe D-evq'thro-o-g~lecto-octonicphenylhydrazide was obchemical oxidation thus appears t o be the enantiomorphous tained by heating 0.1 g. of lactone and 0.1 ml. of phenylL-glycero-D-nznnno-octulose ( I X ; see discussion in the first part of this paper) and a definitive proof of its structure will hydrazine in 1 ml. of water for a few minutes on the steambath. The mass solidified, and the product was purified be sought when more of the octitol V I becomes available. Syntheses. Pyridine Rearrangement of D-threo-L-gahcto- by two recrystallizations from water as clusters of tiny needles. The phenylhydrazide melted a t 19:3-195" :inti Octose (X), Reduction of the Products, and Isolation of Lf23" in water ( e 0.2). erythro-L-galacto-Octitol and Its 0ctaacetate.-A 5.8-g. showed [.Y]~OD A n d . Calcd. for Ci4H24N20s: C, 48.5.5; H , 6.40; S, sample of D-threo-L-galacto-cctose (X) monohydrate1s was 8.09. Found: C, 48.62; H , 6.41; N, 7.87. heated at 100' in vacuo for- 3 hours and the dried octose refluxed in 45 ml. of dry pyridine for 9 hours. The solution was Reduction of D-erythro-D-galacto-Octonic ?-Lactone (XIV) concentrated and shown to contain octulose by paper chro- to D-ei.ythro-D-galacto-Octitol (VI).-Reduction of 0.33 g. Of matography (see preceding paragraph). The mixture of the lactone with 0.5 g. of sodium borohydride according to sugars was dissolved in 70 ml. of water and reduced with 1.5 the directions of Wolfrom and Wo0d3~yielded an octitol and g. of sodium bcrohydride. -4fter removal of sodium ions and some octonic acid. The latter was removed on a column boric acid in the usual way, i.e., by Amberlite IR-120 and of Amberlite IRX-400 ion-exchange resin. The octitol ohevaporation with methanol, respectively, the mixture of tained on concentration was recrystallized from ethanol consolid alditols was shaken with 50 ml. of water. The insoluble taining a small amount of water. It weighed 0.15 g., portion (2.5 g.) was recrystallized from hot water and identimelted at 170-171° both alone and in admixture with the fied by melting point and mixed melting point of 231' as the naturally occurring D-erythro-D-galncto-octitol monohydrate, known D-thi.eo-L-galacto-octitol.'8 The aqueous extract, and showed [ c x ] ~ O D -12' in 5% ammonium molybdate ( e upon concentration, yielded an additional 1.1 g. of the same 0.41) and +62" in acidified molybdate ( c 0.33). The natalditol. Finally, the mcther liquor yielded 0.36 g. of the es- ural and synthetic octitols migrated at the same rates on pected epimeric alditol melting a t 150-160°. This was paper chromatograms run in all four solvent systems. Acetidentified as D-thveo-L-kdo-octitol, which can also be named ylation of the synthetic octitol yielded an octaacetate that L-er>~thro-L-galacto-octitol and is then readily recognized as melted at 99-100", both alone and when mixed with the octathe enantiomorph of the naturally occurring octitol V I . acetate from the octitol from the avocado, and showed Paper chromatograms run in solvent .4 for 48 hours and in [ C X ] ~ ' ~+Dl o in chloroform (c 1.0). Infrared spectra of the solvent C for 18 hours showed identical rates of migration two octaacetates in chloroform solution were identical. for the enantiomorphs. Upon acetylation the synthetic D-erythro-D-galacto-octose (XV).-The reduction of 4.4 g. enantiomorph yielded 0.16 g. of large prisms that, after of D-erythro-D-gUlUCtG-oCtoniC y-lactone (XIV) was carried recrystallization from cthanol, melted a t 97-98' and showed out in ice-cold sodium hydrogen oxalate buffer solution by the general procedure of I ~ b e l l . 3 ~A total o f 400 g. of 2% [ a l z 0-~I o in chloroform (c 3) as compared with m.p. 99100" and [.]O' D +2" for the octaacetate of the naturally oc- sodium amalgam was added in 50-g. portions a t intervals of curring octitol described earlier in this paper. The infrared one hour while a pH of about 3 was maintained by the addispectra of the two octaacetates in chloroform solution were tion of oxalic acid dihydrate to the vigorously stirred soluidentical. Upon deacetylation of the synthetic product, tion. The next morning the mixture was filtered and an the L-erythro-L-galncto-octitol monohydrate melted a t 168additional 250 g. of amalgam was added in the same manner 169' and showed [a]*OD - 70" in acidified molybdate solu- to the filtrate. The mixture was again filtered, a considertion (c 0.16) as compared with m.p. 169-170" and [ a l z o ~able amount of the dissolved salts was precipitated by the addition of 3 liters of ethanol, and the remainder by deioniza+66" for the enantiomorph, D-evythro-D-gnlacto-octitnl tion with Amberlite IR-120 and Duolite A-4 resins. Con(VI). Synthesis of D-erythro-D-gnlarto-Octonic Acid (XII) from centration of the solution left 3 g. of sirup that contained D-glyrero-D-inUnno-HeptOSe (XI) .-Twelve grams of Dsome unchanged lactone; the latter was removed by progl2'cero-a-D-malznc-heptose hexaacetateZ0 was deacetylated longed shaking of an aqueous solution of the sirup with Duocatalytically with sodium methoxide and the sirupy heptose lite A-4 resin. Concentration of the solution yielded 0.8 g. treated with 1.8 g. of sodium cyanide by Hudson's modifica- of a sirup that crystallized readily on addition of methanol. tiorP of the cyanohydrin synthrsis. .ifter saponification Upon recrystallization from methanol containing a small of the nitriles, removal of the sodium ions and concentraamount of water the octose separated as acicular prisms that tion, the 5.9 g. of product was obtained as a solid residue. melted at 174-175". The rotation changed from [alZoD \\'hen this material was dissolved in 25 ml. of hot water i t f 4 5 " (after a few minutes) t o +64" (overnight, constant) in deposited 1.3 g . of elongated prisms of D-er.ythro-D-gn/nc-/o- water (c 0.72) and the octose accordingly is to be designated octcjnic acid innnohydrate; once recrystallized from water it D-evythro-P-D-galacto-octose. I n relation t o galactose, the melted at 137'' t o a thick sirup and showed [ C X ] * ~ D-7' in octose showed Rgal 0.50 in solvent A and 0.46 in solvent B. water ( c 1). \\-hen heated for 5 hours a t 100' in Z ~ C Z I the O On paper chromatograms i t was detected as a dark green mnrioliydrate lost about one molar equivalent o f water spot with the p-anisidine hydrochloride spray reagent. (calcd. 6.57, found 5.91) and was converted to the anhydrous Anal. Calcd. for C8H1608: C, 40.00; H, 6.71. Found: acid rather than to the lactone; a test for lactone mas negaC, 40.26; H, 6.98. tive when a spot of the material on paper was sprayed with Pyridine Rearrangement of o-erythro-D-ga~acto-octOSe alkaline hydroxylamine followed by ferric chloride. (XV).-A small sample of t h e octose was rearranged by boilAnal. Calcd. for CsH160U.H20,C, 35.04; H , 6.62. ing in dry pyridine for 8 hours. The solution then showed Found: C , 35.24; H , 6.40. Calcd. for CIHIGOY:C, 37.50; the presence of an octulose t h a t moved a t the same rate as H,6.29. Found(driedat 10Ooinwicuo): C,37.71; H,6.51. the avocado octulose when paper chromatograms were run D-Evythi.G-D-ga~ncto-Octonic -/-Lactone (XIV) and Dfor extended periods of time in all four solvent systems. erythuo-D-galacto-Octonic Phenylhydrazide.--\Vhen 0 . 2 g. (31) hl. L. Wolfram and H. B. Wood, Jr., ibid.. 73, 2933 (1951). of the octonic acid monohydrate was dissolved in 40 ml. o f (32) H. S. Isbell, U.S. P a t e n t 2,632,005, March 17, 1953; see also boiling 2-ethoxyethanol, refluxed for 5 hours, and the solH. S. Isbell, J. V. Karabinos, H. L. Frush, N. B. Halt, A. Schwebel vent removed by distillation, the solid residue was found to (30) C . S. Hudson, THISJOURNAL, 73, 4498 (1951); J. W. Pratt nnd S. K. Richtmyer, ibad., 77, 1006 11955).

and T. T. Galkowski, J. Reseavch Natl. B U Y .Standards, 48, 163 (1962), a n d J. W. Pratt, N. K. Richtmyer and C.S.Hudson, THISJOURNAL, 75, 4503, footnote 20 (1953).

Products from Sedum Species .-Isolation.-Preliminary experiments were carried out on an aqueous extract from 2600 g. of Sedum Adolphi Hamet. Gums were precipitated with methanol, and D-glucose, D-fructose and sucrose were removed by fermentation with bakers' yeast. Deproteinization, deionization, and concentration left 5.5 g. of a dry sirup. X paper chromatogram run in solvent C for 40 hours and sprayed with orcinol-trichloroacetic acid revealed a strong blue sedoheptulose spot; a very weak blue spot just above i t corresponding to the second anhydride of sedohepand tulose, namely, 2,7-anhydro-(3-~-altvo-heptulofuranose~~; a weak blue spot just below it corresponding to sedoheptulosan (2,7-anh;~dro-(3-~-altuo-heptulopyranose).Below the sedoheptulosan spot was a weak gray spot, believed t o be an octulose, and below that a weak, elongated, bluish spct (or probably, as the ammoniacal silver nitrate spray indicated, 3 separate spots that appeared as a single long spot with the orcinol reagent). -12-year-old extract of Sedum spectabile that had been treated similarly showed the same 5 spots with orcinol-trichloroacetic acid. -iccordingly, a large-scale operation was begun in order to study further the presumed octulose from Sedum species as well as the substances responsible for the still slower moving spot or spots that appeared with the orcinol spray. For this purpose we had available the extracts, obtained over a 4-year period, from a total of 385 kg. of Sedum plants (principally S. spectabile). These had been precipitated with methanol t o remove gums and mere then treated with yeast to remove fermentable sugars. The resulting solutions were deproteinized, deionized, and heated in 0.2 N sulfuric acid a t 40' to convert about 907" of the sedoheptulose to sedoheptulosan .33 After deacidification with Duolite -1-4 and concentration, about 2100 g. of sedoheptulosan hydrate was recovered by crystallization. The mother liquor was retreated with acid, and an additional 300 g. of sedoheptulosan hydrate obtained. On a paper chromatogram sprayed with aniline hydrogen phthalate the mother liquor appeared to contain about seven aldoses, and so it was oxidized with bromine water in the presence of a considerable excess of calcium carbonate. The solution was deionized and concentrated to a sirup that weighed 345 g.; calculations showed that about 100 g. of aldoses had been removed by the hypobromite oxidation. The sirup, chromatographed on paper in solvent C and sprayed with aniline hydrogen phthalate, showed negligible amounts of aldoses; with orcinol-trichloroacetic acid it showed the five spots representing ketoses and their non-reducing anhydrides; and with ammoniacal silver nitrate it showed about a dozen separate spots that included the ketoses, their anhydrides and polyhydric alcohols. Fractionation on Cellulose.-The 345 g. of sirup obtained as described above was subjected t o chromatography on a cellulose column 52 cm. high and 7.5 cm. in diameter. The sirup was divided into five portions varying from 55 t o 88 g., and ea.ch portion was put on the column and eluted with mixtures of benzene and methanol containing 170water, starting with 5 : 1 ( +1% H?O) and gradually decreasing the proportion of benzene until a 1: 1 ( 1yc HzO) ratio was used for the final elution. The many samples of eluate from the five runs were examined by paper chromatography and eventually distributed among eleven fractions, of which the first 4 have not yet been investigated. Fraction 5 contained principally sedoheptulosan, of which 50 g. has so far been obtained crystalline. Fraction 7 yielded D-mannitol and fraction 11 a considerable quantity of myo-inositol. Fractions 8-10 contained the octulose, and fractions 9-11 contained the material that appears as the slow-moving, elongated, blue spot on a paper chromatogram sprayed with orcinol and trichloroacetic acid. The sedoheptulose and its anhydrides, represented by blue spots that moved faster than the octulose, appeared in fractions 4-9. Because the separation of the large amount of complex mixture under the conditions described above was not clean cut, a 3.6-g. portion of fraction 9 was removed for the study reported in the following section, and the remainder of fractions 8-11, weighing 33.5 g., is currently being investigated by fractionation with other solvent systems. Examination of Fraction 9. Isolation of D-glycero-Dyluco-Heptitol (p-Sedoheptitol, V).-From the first 115 g. o f sirup that was fractionated on the large cellulose column

(representing about 130 kg. of Sedum leaves and stems) was obtained 3.6 g. of sirup, called fraction 9, that seemed to contain the largest proportion of octulose. This sirup was refractionated on a smaller cellulose column (325 mm. high and 50 mm. in diameter) using 1-butanol half saturated with water until the eluate showed only a trace of octulose, and then changing to 1-butanol three-quarters saturated with water. Seven fractions were obtained, with fractions 9C and 9D showing only the gray, octulose spot and fractions 9E and 9 F showing only a slower-moving blue spot on paper chromatograms sprayed with orcinol-hydrochloric acid. Fraction 9C (1.68 g.), when run on a paper chromatograni in solvent B for 2 days and sprayed with periodate-per1.2) and another manganate, showed the octulose spot (RBUDr spot above it. The sirup was dissolved in a mixture of hot methanol and ethanol and the solvents allowed to evaporate slowly a t room temperature. After a few days the crystals that had deposited were filtered and washed with methanol. The 0.4 g. of product melted a t 131-132" and was identified as 8-sedoheptitol by mixed melting point (no depression), identical Rr-values on paper chromatograms run in all 4 solvents, and identical infrared spectra on direct comparison with authentic material, and by its rotation of [ C Y ] ~ "+46" D in 570 ammonium molybdate as compared t o the value of $49.6" reported in the l i t e r a t ~ r e . ~ ~ Isolation of D-g2ycel.o-D-manno-Octulose (VII).-The remainder cf fraction 9C was combined with 9D (0.1 g.) and some additional ,p-sedoheptitol removed by crystallization. The residue (ca. 1 g.) was put on the small cellulose column and eluted with 1-butanol one-quarter saturated with water. Eight fractions were collected, with 0-sedoheptitol being detectable by paper chromatography only in the first two fractions. The combined middle fractions 4-6 (0.16 9.) showed [ c Y I ~ O D $20" in methanol ( c O.Bj, the same value that had been obtained for the sirupy octulose from the avocado. On paper chromatograms run in all four solvent systems the Sedum octulose had the same mobility as the avocado octulose. Infrared spectra of the thoroughly dried octulose sirups as films from methanol were identical. A 16-mg. sample of the Sedum octulose was oxidized with 2 molar equivalents of lead tetraacetate as described for the avocado octulose. The resulting pentose in this case also had the same mobility as ribose when run on paper chromatograms in all four solvent systems. Another small sample was degraded with oxygen in cold alkaline solution; paper chromatograms run in solvents B and D indicated t h a t the principal product here, as from the avocado octulose, was Dglycero-D-manno-heptonic lactone. Preparation of D-glycero-D-manno-Octose Phenylosazone from the Sedum Octulose, the Avocado Octulose and Derythro-~-galacto-Octose.-il 0.10-g. fraction of octulose sirup from Sedum, containing no detectable amount of 0sedoheptitol or other impurity, was dissolved in 3 ml. of water, together with 0.2 g. of phenylhydrazine hydrochloride and 0.3 g. of sodium acetate trihpdrate, and the mixture heated for 30 minutes in a hot water-bath a t 95". Aiyellow osazone separated and, after cooling, it was centrifuged, and washed well with water and then ether. On recrystallization from 5070 aqueous ethanol it separated as needles; yield, 10 mg. The phenylosazones were prepared in the same way from the sirupy avocado octulose and from the synthetic D-erythro-D-galacto-octose ( X Y ) . Each of the phenylosazone samples showed the same melting point, 188189" with decomposition, with no significant depression of that value when mixed melting points were made. X-Ray powder diffraction patterns and infrared spectra in potassium bromide pellets were identical for all three phenylosazones. Anal. Calcd. for CzOHZ6~4O6: C, 57.40; H, 6.26; S, 13.39. Found (from Sedum octulose): C, 57.11; H , 6.30; N, 13.24. Found (from synthetic octose): C, 57.39; H, 6.10; N, 13.20.

(33) X. R. Richtmyer and J. W, Pratt, THISJOURNAL, 78, 4717 (193G1.

( 3 4 ) A. T.Merrill, W, T. Haskins, R. M. I-Iann and C. S. Hudson, {bid.. 69, 70 (1917).

+

Acknowledgments.-The authors wish to thank Dr. William C. Alford and his associates of the Analytical Services Unit of this Laboratory for obtaining the infrared spectra and the elemental analyses; Dr. Alford and also Mr. Charles G. Remsburg, formerly of this Laboratory, for their

3434

ROBERT K . NESSAND HEWITTG. FLETCHER, JR.

contributions of Sedum spectabile; Mr. John T. S p e s and Mr. Edward LV. Tracy of this Laboratory for their help in processing the Sedum; Dr. John C. Keresztesy and his associates, Dr. Norman E. Sharpless, and Dr. Laura C. Stewart, of other laboratories of this Institute, for help in concentrating the Sedum extracts, obtaining the X-ray powder

Vol. 82

diffraction patterns and with the Acetobacter suboxydans study, respectively; and bir. W. E. Clarke, 111, of Lynchburg, Va., for his collaboration in the study of Sedum Adolph; while he was a participant in the High School Science Teacher Study-Work Institute during the summer of 1958. BETHESDA 14, MD.

[CONTRIBUTIONFROM THE XATIONAL INSTITUTE O F ARTHRITIS AND METABOLIC DISEASES,NATIONAL INSTITUTES O F HEALTH]

2-Deoxy-D-ribose.

1V.l A Direct Synthesis of Z’-Deoxyadenosine and its Anomer through 2-Deoxy-D-ribose Derivatives2 BY ROBERTK. NESSAND HEWITTG. FLETCHER, JR. RECEIVED JANUARY 11, 1960

Mono-p-nitrobenzoylation of 2-deoxy.~-ribose diisobutyl dithioacetal affords a 5-0-p-nitrobenzoyl derivative. Demercaptalation of this, followed by further p-nitrobenzoylation, gives the two anomeric 2-deoxy-~-ribofuranosetri-p-nitrobenzoates. Conversion of these esters t o amorphous 2-deoxy-3,5-di-0-p-nitrobenzoyl-~-ribosyl chloride, followed by condensation with chloromercuri-6-benzamidopurineand removal of the protecting groups, has led to the isolation of 2’-deoxyadenosine, its anomer 9-( 2-deoxy-or-~-ribofwanosyl)-adenine and a third nucleoside, probably 7-(2-deoxy-~-n-ribofuranosyl)adenine. A direct preparation of 5-0-trityl-2-deoxy-a-~-ribosefrom 2-deosy-D-riboseis described.

Chemical syntheses of nucleosides containing the 2-deoxy-~-ribofuranosyl moiety have heretofore relied on indirect pathways, preformed pentofurano~ylpyrimidines~ and purines4 being reduced a t carbon two of the sugar moiety through various ingenious transformations. In view of the relative accessibility of 2-deoxy-~-ribose,~ the direct synthesis of such nucleosides through suitably substituted 2deoxy-D-ribofuranosyl halides offers a number of attractive features. The present paper will describe such a direct synthesis! Zinner and co-workers7 have shown that pentose dialkyl dithioacetals may be readily p-nitrobenzoylated a t carbon five and that, through demercaptalation, 5-0-acylpentose dialkyl dithioacetals may be converted to 5-0-acylpentofuranoses. This method of obtaining pentofuranose derivatives was used in the course of the present research. 2-Deoxy-~ribose was converted to its known diisobutyl dithioacetals and mono-p-nitrobenzoylated t o yield the crystalline 5-0-fi-nitrobenzoyl derivative I. Demercaptalation with mercuric chloride and mercuric oxide then gave an amorphous 2-deoxy-5-0p-nitrobenzoyl-D-ribose (11) which was further pnitrobenzoylated, affording the two anomeric 2deoxy-D-ribofuranose tri-p-nitrobenzoates (111 and IV). Either of these esters or a mixture of the two was dissolved in methylene chloride and treated with a (1) 2-Deoxy-D-ribose. 111. C. Pedersen, H. W. Diehl and H. G. 82, 3425 (1960). Fletcher, Jr., THISJOURNAL, (2) For a preliminary communication describing this work see R . K. Sess and H. G. Fletcher, Jr., i b i d . , 81, 4752 (1959). (3) D. M. Brown, D. B. Parihar, C. B. Reese and A. Todd, J. Chem. Soc , 3035 (1958); G. Shaw and R. M. Warrener, ibid., 50 (19Z9). (4) C. D. Anderson, L. Goodman and B. R. Baker, THISJOURNAL, 81, 3967 (1959). ( 5 ) H. W. Diehl and H. G. Fletcher, Jr., Biochem. Preparafions, 8 , in press (1960); Arch. Biochem. Biophys., 7 8 , 386 (1958). (6) Subsequent t o the completion of this work we were informed of the successful synthesis of several pyrimidine 2’-deoxynucleosides by a direct procedure which has some features similar in principle t o t h a t described here; see M. Hoffer, R. Duschinsky, J. J. Fox and N. Yung, THISJOURNAL, 81, 4112 (1959). (7) H. Zinner, K. Wessely, W. Bock, K. Rieckhoff, F. Strandt and W. Nimmich, Ckem. Ber., 90, 500 (1957). (8) H. Zinner, H. Nimz and H. Venner, Chcm. Ber., 90,2696 (1957).

very slight excess of hydrogen c h l ~ r i d e . ~p-Nitrobenzoic acid, which was precipitated in nearly quantitative yield, was removed, leaving a solution of 2deoxy-3,5-di-0-~-nitrobenzoyl-~-ribosyl chloride (V) .lo The amorphous, solvent-free halide was then condensed with chloromercuri-fi-benzamidopurine in dimethyl sulfoxide solution. Protecting groups were stripped off with alkali and the product chromatographed on a cellulose column. 2’-Deoxyadenosine (VI, 9-(2-deoxy-~-~-ribofuranosy~)adenine) was isolated in 10% yield. Its melting point, rotatory dispersion and ultraviolet absorption spectrum clearly showed i t to be identical with the natural nucleoside. A second crystalline nucleoside, isomeric with the natural product, was isolated through chroniatography in 19% yield. In contrast with the first product, i t was markedly dextrorotatory in the wave length range measured (559 to 340 mp). I n the ultraviolet i t showed an absorption peak a t 260 mp characteristic of a 9-substituted adenine.” Hydrolysis with weak acid revealed only adenine, 2-deoxy-~-ribose and unchanged nucleoside. I t appears highly probable that the substance is, therefore, the anomer of the natural nucleoside, namely, 9- (2-deoxy-a-~-ribofuranosyl) -adenine (VII). Finally, the chromatography yielded a small amount (1%) of a third, isomeric nucleoside. From the dextrorotation of this substance and its absorption a t 273 mp (characteristic of 7-substituted adenines”) i t may tentatively be assigned the structure 7-(2-deoxy-cu-~-ribofuranosyl)-adenine. Zinner, Nimz and V e n n e P have described an indirect method of preparing 2-deoxy-5-0-trityl-aD-ribose; in the course of the present work the same substance was prepared rather more conveni(9) This technique is similar t o t h a t employed b y W. W. ZorSO, 5564 (1958). bach and T.A. Payne, Jr., TEE JOURNAL, (10) One of the anomeric forms of this substance has since been obtained in crystalline form; i t and some of its reactions will be described in a future communication. (11) J . 11.Gulland and L. F. Story, J. Chem. Soc., 259 (1938). (12) H. Zinner, H. Nimz and H. Venner, Chcm. Ber., 91, 638 (1958).